The disclosed embodiments of the present invention relate to processing a video input which transmits pictures of different views, and more particularly, to a method of processing a video input by detecting if a first picture of one view is correctly paired with another picture of another view for specific presentation time and related processing apparatus thereof.
With the development of the science and technology, users are pursuing three-dimensional (3D) and more real video outputs rather than high quality two-dimensional (2D) video outputs. The 3D video content may be encoded and transmitted by one of a variety of formats. For example, when a side-by-side format or a top-and-bottom format is employed, one encoded video frame would include a left-eye picture and a right-eye picture arranged therein. Thus, with regard to decoding of the encoded video frames having the side-by-side format/top-and-bottom format, the left-eye pictures and right-eye pictures would be produced in pairs in the time domain before displayed.
In a case where the multi-view video coding (MVC) is employed to encode the 3D video content, the left-eye picture and the right-eye picture are arranged in different encoded video frames. For example, a multi-track storage scheme may be used for storing encoded video frames having the MVC format. Therefore, encoded video frames which transmit left-eye pictures are stored in one track, and encoded video frames which transmit right-eye pictures are stored in another track. Thus, with regard to decoding of the encoded video frames having the MVC format, it is not guaranteed that the left-eye pictures and right-eye pictures would be produced in pairs in the time domain before displayed.
Due to certain errors happening at the encoder side and/or the decoder side, one left-eye picture included in a decoder output may not be correctly paired with one right-eye picture included in the decoder output for following playback, and vice versa. For example, one left-eye picture and one right-eye picture generated in a pair due to decoding of the encoded video frames may have different presentation time stamps (i.e., different playback time points). Besides, it is also possible that one left-eye picture included in the decoder output is not paired with any right-eye picture, and vice versa. Therefore, at one particular presentation time, only one of the left-eye picture and the right-eye picture is available in the decoder output. These pairing errors existing in the decoder output may greatly degrade the 3D video playback quality, leading to uncomfortable 3D visual experience for the viewer.